What’s remarkable about @typesafeai’s Jev isn’t just how obvious it feels in hindsight, it’s that it didn’t come from a major lab. It’s an inspiring glimpse of AI’s future: an explosion of frontier innovation from teams across the industry, not just a handful of labs.

The world needs more vertical benchmarks that aren’t exposed in training data. Today we’re launching Accounting Bench focused on real world accounting tasks. What’s notable is even Fable and Astra can only fully solve ~20% of the tasks. There’s a lot of room to run.

When I first started working in ML/AI, the idea of exponential progress felt like fantasy. Frankly, I didn’t buy into it until very recently. But when I postmortem where I was wrong, it feels obvious: the curve of human progress follows this compounding slope. The last 100 years alone are indicative of the insanity. And when you think about humanity becoming a multiplanetary species, achieving abundance, and eradicating disease of course it all happens through the exponential. It almost certainly has to. The gravity against progress is so severe that escape velocity becomes the natural extension. Recently, I think many of us feel like we’re peering through the doorway into the future that isn’t too far away anymore. And through it somehow we’re more connected to the nature of the universe.

I resigned from @tryramp today. I spent the last year there doing applied AI work and neither CEO is acting responsibly. They are racing toward self-closing books with no clear plan for what happens when companies have saved too much time and money. More thoughts below.

I hear this the most with family/people I know outside of tech It often feels like we’re living in completely different planes of existence The next few years are going to be incredibly disorientating and scary for the average American